2011 Geoarchaeological Research at St. Catherines Island 79

CHAPTER 3 GEOARCHAEOLOGICAL RESEARCH AT ST. CATHERINES ISLAND: DEFINING THE GEOLOGICAL FOUNDATION Gale A. Bishop, Brian K. Meyer, R. Kelly Vance, and Fredrick J. Rich

St. Catherines Island is located at the head Understanding ’s sea level changes of the Georgia Bight, midway in the string of 12 demands accommodation of known data that barrier islands forming the Golden Isles of the constrain models and resultant sedimentologi- Georgia coast (fig. 3.1; Foyle, Henry, and Al- cal effects on shoreline position and elevation exander, 2004). With no proximal source of flu- (Leatherman, Zhang, and Douglas, 2000; Doug- vial sediment, St. Catherines Island is dependent las, Kearney, and Leatherman, 2001; Coe, 2003). upon net longshore transport of sand from north The height of maximum sea level rise in Georgia to south along the Georgia coast (Hails and Hoyt, is equivalent to the elevation of the highest coast- 1969; McClain, 1980; Clayton et al., 1992). Net al deposits of the Wicomico Shoreline, or ter- southward longshore transport is indicated by the race. Although the array of preserved Pleistocene chenierlike Delta with south- shoreline deposits or terraces provides evidence ward accretion (Alexander and Henry, 2007), the of progressive lowering of sequential sea level chenierlike Altamaha Delta with its southward highstands (Stapor and Mathews, 1983; Gayes et accretion, and southward migration of islands al., 1992), it says little about the sea level low- along the Georgia coast, and the building of the stands during glacial stages. Vertebrate fossils cape at Cape Canaveral, Florida (Davis, 1994). and archaeological artifacts from subtidal coastal This net southward transport of sediment is inter- environments (DePratter and Howard. 1977, rupted by local effects of flood and ebb currents 1981) and from the continental shelf allow par- at Georgia’s sounds, each forming a horizontal tial reconstruction of lowstands (Garrison, 2006). sand circulation pattern (Oertel, 1972a, 1977; Systematic survey and submarine excavation at Oertel and Foyle, 1993) exchanging sand with Gray’s Reef National Marine Sanctuary, 32 km the shelf and islands themselves (Swift, 1968; (20 mi) offshore Georgia, and nearby J-Reef, in Pilkey, et al., 1981). Interruption of this flow of the Atlantic Ocean, have identified two localities sand by damming rivers to the north and dredg- containing vertebrate fossil remains and two arti- ing the Savannah Ship Channel across the Savan- facts, an organic artifact (a bone antler tool) and nah River Delta (U.S. Army Corps of Engineers, a lithic (a projectile point typologically assigned 1991, 1996), has conspired with rising sea level to the early Middle Archaic period (Garrison, to make St. Catherines one of Georgia’s most ero- 2006). Postglacial sea level must have recovered sional barrier islands (Griffin and Henry, 1984). the 17–20 m (56–66 ft) depths at these locations Recent sea level trends are documented on the during the Archaic period, ~8000 yr b.p., and as National Oceanic and Atmospheric Administra- much as 40,000 yr b.p. prior to that, the areas sur- tion Sea Levels Online site for Fort Pulaski and rounding Gray’s Reef and J-Reef were part of an Savannah (NOAA, 2003), indicating that cur- exposed coastal plain. Relative sea level rose in rently sea level is rising at a rate of 3 mm/yr (± the Holocene (post-12,000 yr b.p.) continuing to of 0.2 mm/yr; 1935–1999). At Fernandina Beach, rise from full post-Wisconsin lowstand of over Florida, the rate is 2 mm/yr (1897–1999). 100 m (328 ft) below present sea level (Garri- 80 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94

A GEORGIA B SOUTH CAROLINA SOUTH WILMINGTON ISLAND CAROLINA SKIDAWAY ISLAND Tybee Island Penholoway Fm. Wassaw Island Ossabaw Island

Pamlico Pr. Anne Fm.Holocene GEORGIA Formation ST. CATHERINES Wicomico Formation ISLAND MIOCENE ATLANTIC OCEAN Blackbeard A Island Index Sapelo Cross- Island Section A' (Below 1C) Little St. Simons Island St. Simons ATLANTIC Penholoway Fm. Island Wicomico Fm. LEGEND OCEAN

Silver Blu Fm. Jekyll Fossil Island Callianassa LEGEND Pr. Anne Fm. Holocene Talbot Fm. burrows Holocene barrier Barrier island island

SWAMP GA facies Pleistocene Cumberland barrier island OKEFENOKEE Lagoonal Island marsh (Silver Blu ) FLORIDA facies FLORIDA SCALE SCALE 0 10 20 30 0 5 10 km mi C WEST EAST A A' FT.MSL FT.MSL

120 Wicomico Formation Penholoway 120 Formation Talbot Silver 80 Fm. Pamlico Blu 80

Formation Princess Fm. Anne Formation 40 Holocene 40 0 sea level 0 Miocene -40 -40 LEGEND Pliocene Pleistocene -80 -80 Barrier island SCALE facies 0 2 4 6 8 10 Lagoonal marsh KM facies

Fig. 3.1. Development of successive shorelines on the Georgia coast: A, successive shorelines, headlands, and intervening marshes (formations); B, recent Silver Bluff Pleistocene and Holocene shorelines of Georgia forming modern Golden Isles; C, cross section of Pleistocene to Holocene sediment veneer of the Georgia coastal plain (after Hails and Hoyt [1969] and Hoyt and Hails [1967]). 2011 Geoarchaeological Research at St. Catherines Island 81 son, 2006). Sea level had dropped prior to the sulted in deposition and erosion of a seaward-dip- Late Glacial Maximum ~21,000 yr b.p. (Marine ping veneer of Pleistocene sediment arranged in a Oxygen Isotope Stage-2 [MIS–2]). Differential series of barrier island sequences that are younger elevations of ancient barrier island or shoreline to the east. The deposition of coastal terraces or complexes and structural evidence suggest that barrier island ridges (Wicomico, ~29–30 m [~98 tectonic as well as eustatic controls have been ft]; Penholloway, ~23 m [~75 ft]; Talbot, ~12–14 in effect. Superimposed upon these eustatic and m [~39–46 ft]; Pamlico, ~8 m [~26 ft]; Princess tectonic effects are sedimentological pulses pro- Anne, ~4.5 m [~14 ft]; Silver Bluff, ~1.5 m [~5 duced by Pleistocene climate changes and evolv- ft]; and Holocene) in Georgia form a continuous ing physical conditions influenced by possible veneer of Pleistocene sediment of varying thick- coastal plain stream capture of inlets and sounds ness and lithology (Huddleston, 1988). Shoreline along the Georgia coast (Chowns et al., 2008; elevations were based on the elevations of fossil Chowns, chap. 9, this volume). burrows of Callichirus major (Say, 1817–1818; The profound effects of sea level change were Rodrigues, 1983). emphasized in a figure of Pleistocene coastal Hoyt and Hails (1967), Hails and Hoyt (1969), Georgia by a geographer (LaForge, 1925) in a Pickering and Murray (1976), Linsley (1993), map graphically depicting the configuration of Bishop et al. (2007), Linsley, Bishop, and Rol- southeast Georgia during the Wisconsin high- lins (2008), Reitz (2008), Thomas (2008), and stand. MacNeil (1950) described the shorelines/ others, have suggested that the most recent set terraces from Georgia and Florida citing four of Georgia barrier islands consist of older sedi- marine terraces and shorelines between sea ment deposited about 35,000 to 40,000 years ago level and ~29–30 m (~100 ft) above sea level. with younger sediment to the east accumulat- He proposed that the higher Okeefenokee and ing against the island about 4000 to 5000 years Wicomico shorelines could be correlated with the ago (fig. 3.1, upper right). The Pleistocene parts Yarmouth and Sangamon interglacial stages, re- of the islands formed when the sea level was spectively, the Pamlico Shoreline correlated with ~ 2.0 m (6.5 ft) above the present level, before the a mid-Wisconsin ice recession, and the lowest, formation of the last great continental ice sheet the Silver Bluff Shoreline with post-Wisconsin. of the Pleistocene epoch, the Wisconsin Glacial Hoyt, Henry, and Weimer (1968); Hails and Hoyt Stage, that lowered sea level 80 m (~260 ft), plac- (1969), and Hoyt and Hails (1967) described the ing the shoreline 128 km (~80 mi) offshore near formation of a veneer of Pleistocene sediments the present edge of the continental shelf. across the Georgia coastal plain (fig. 3.1, lower) Gray’s Reef (Henry and Hoyt, 1968; Hunt, as sea level fluctuated throughout the Pleistocene 1974) is an exposed dolomitized micritic lime- and apparently dropped with each subsequent in- stone hardground interpreted by Woolsey (1977) terglacial sea level rise; these sediments built a to be stratigraphically and lithologically equiva- sequence of barrier island complexes (fig. 3.1A, lent to the Raysor shelly sand. It had been origi- B) that got progressively younger, and lower in nally referred to as the Sapelo facies of the Du- elevation, toward the present coastline. plin formation. Subsequent geological studies The problem of sea level rise and fall illus- include Continental Shelf Associates, Inc., 1979; trates the complexity of geology as a unique criti- Henry and Giles, 1980; Henry, 1983; Van Dolah, cal thinking paradigm applying logical reasoning Calder, and Knott, 1984; Van Dolah, Wendt, and first to the stratigraphic problem, then to the in- Nicholson, 1987; Henry, Dean, and Olsen, 1987. terpretation of the stratigraphy (Frodeman, 1995). Huddleston (1988) later revised the lithostratig- Hoyt, Weimer, and Henry (1964) described the raphy of the Georgia coastal sediments. stratigraphy of the mid to late Sangamon and According to this literature, the present glob- Pamlico, Princess Anne, and Silver Bluff paleo- al rise in sea level began approximately 20,000 shorelines on the Georgia coast. The Silver Bluff years ago, moving across the exposed continen- deposits form the core of many modern barriers tal shelves about 1 m/100 yr until ~6000 years including Ossabaw, St. Catherines, Sapelo, St. ago, at which time the rate of rise slowed to ap- Simons, Jekyll, and Cumberland islands with as- proximately 0.3 m/century until today. Garrison sociated Silver Bluff marsh lithosomes currently (2006) documents the existence of Gray’s Reef submerged by the Holocene transgression. Along above sea level some 15,000 years ago when the Georgia coast these sea level fluctuations re- Georgia’s shoreline was more than 60 mi east 82 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 of its present position. Off the coast, divers have 1 m/yr over the past century, leading to stable discovered fossils of now-extinct land-dwelling central shorelines with most fluctuations occur- animals, such as ground sloths, mastodons, and ring adjacent to tidal inlets. In Georgia, barrier early camels, horses, and bison. Gray’s Reef islands commonly exhibit lateral accretion, pro- inundation began ~7000 yr b.p. (Henry, 2005). gradation, and dynamic equilibrium, which are Thus, after several cycles of submergence and the primary responses of barrier coastlines. emergence, with the latest period of exposure Riggs and Cleary (1993) recognized that many lasting 40,000 years, the substrate was once East Coast barrier islands are “perched” barriers again covered by the ocean and once again be- whose forms are strongly determined by ante- came a live bottom. cedent topography and the types of antecedent Swift et al. (1972) recognized that inlet, ebb- sediments available to wave/current erosion on tidal delta, and estuarine deposits from low- the shoreface and foreshore. The structural and stands of late Pleistocene and Holocene sea level stratigraphic characteristics of a barrier island are principal sources of sediments for building complex influence barrier island morphology, current beaches. Chester DePratter and James inlet development, and beach dynamics (Hoyt, Howard (1977, 1981) studied the “History of 1967). Scott, Gayes, and Collins (1995) and Scott shoreline changes determined by archaeological and Collins (1995) recognized a rapid increase in dating: Georgia coast,” documenting the exis- sea level rise to a point about 1.5 m above cur- tence of intertidal archaeological sites (see also rent sea level in South Carolina. Wanless (2002) Caldwell, 1971). described Holocene coastal change on the mud- Winker and Howard (1977) described diffi- dominated microtidal mangrove coast of west culties in correlation of Pleistocene paleoshore- Florida that is migrating landward at rates of as lines of the lower coastal plain between Florida much as 4 m/yr with a 3 mm/yr relative sea level and Virginia and presented a new terminology rise rate (causing coastal submergence). (Chatham, Effingham, and Trail Ridge sequenc- Langley et al. (2003) described the quan- es) for paleoshorelines. Well-developed trellis- tification of shoreline change on the Georgia style drainage networks were described land- coast on Wassaw and St. Catherines islands us- ward of the Talbot paleoshoreline and dendritic ing shoreline data from 1856 and 1924 and re- drainage patterns prevail at lower coastal plain viewed the history of shoreline mapping on the elevations associated with the modern through Georgia coast. The authors state that Wassaw and Pamlico-Talbot paleoshorelines. They suggested St. Catherines islands exhibit shoreline-change that up to 50 m of downwarping and upwarping patterns proposed by Hayes (1994) to be typical along the Orangeburg Trail Ridge scarp between of Georgia Bight barriers in that erosion occurs north Florida and southern North Carolina was on the updrift ends and accretion occurs on the possible. Pilkey et al. (1981) described common downdrift ends. While this is a good general- lagoonal deposits on the shelf and proposed that ization, it is an oversimplification, being true at southeastern barrier islands are not just mid- to limited temporal and spatial scales (Foyle, Hen- late-Holocene features but migrated across most ry, and Alexander, 2004). Langley et al. (2003) of the shelf during the current transgression. Ma- show that St. Catherines Island does not exhibit son (1993) determined the usefulness of beach this typical behavior. The data appear to reveal ridge archaeology for identifying rates and tim- the downdrift migration of an accretionary bulge ings of coastal change, climate change, and sea that is probably fed by onshore-migrating sand level variation for progradational coasts based on bars from the updrift ebb-tidal delta. However, the fact that human settlement favors open coast- the time frame within which this process typi- lines that correlate with changes in the shoreline. cally operates on the Georgia Bight is not well Gayes et al. (1992) identified a mid-Holocene (4.2 known. St. Catherines Island has become shorter ka) highstand of relative sea level at Murrells In- (1852/1871–1911/1924) and shows shoreline re- let, South Carolina followed by a fall in sea level treat along its entire length with some stability of 2 m until 3.6 ka and then a constant sea level along its central portion (Goodfriend and Rollins, rise of 10 cm/century to the present. McBride and 1998). Griffin and Henry (1984) attributed this Byrnes (1993) compared Louisiana, Mississippi, to the island’s distal location downdrift of sig- and Georgia barrier coastlines. Shoreline change nificant sediment input from the Savannah River. rates on the Georgia coast have averaged about Shoreline retreat rates on St. Catherines Island 2011 Geoarchaeological Research at St. Catherines Island 83 vary from –1.6 to –9.2 m/yr (see Bishop and fluvial sand from the short mainland creeks en- Meyer, chap. 14: fig. 14.11; and Potter, chap. 7). tering St. Catherines and Sapelo sounds. These Fred Pirkle led the study of the geological his- conditions have made St. Catherines Island one tory of the Georgia coast to better define and lo- of Georgia’s most dynamic and erosional barrier cate deposits of heavy minerals (Smith, Pickering, islands (see fig. 3.3). In this context, St. Cath- and Landrum, 1968; Pirkle et al., 1991; Pirkle, erines is a sentinel island for the other barrier Pirkle, and Pirkle, 2007) and define their mode islands of the Georgia coast—predicting increas- of accumulation as placers left behind on the ingly erosional conditions for all the Golden Isles backbeach as swash winnows out the less dense as these conditions continue and strengthen. quartz fraction. Bishop studied the distribution of Erosion of the east and north shores was doc- heavy minerals on St. Catherines Island (1990), umented in the 1970s by McClain (1980). Morris finding them concentrated along the backbeach and Rollins (1977) mapped biological associa- in a series of nodes, in the backbeach dune fields, tions within relict marsh mud exposed as erosion and onto the midbeach. Vance and Pirkle (2007) exposed ancient marshes along the beaches of summarized the distribution and provenance of St. Catherines Island. Monitoring erosion and heavy minerals on the Georgia coast. accretion of the shoreline at St. Catherines was A burst of geological research in the late expanded during the past 20 years by students 1990s resulted from the Georgia State University in The University of the South’s Island Ecology master’s thesis work of Robert Booth, under the Program to include more than 25 stations with supervision of Fred Rich. Booth et al. (1998), more or less continuous measurement of shore- Booth, Rich, and Bishop (1999), and Booth et al. line change. Bishop initiated documentation of (1999) described aspects of the stratigraphy of beach habitat deterioration through an annual St. Catherines Island. Booth and Rich (1999) de- Rapid Habitat Assessment (Bishop and Marsh, scribed a dense peat from the Cracker Tom bridge 1999b) survey to characterize those deteriorating core at a depth of 5.02–5.12 m (total depth) that conditions through a semiquantitative of sea tur- consisted of 85% monolete pteridophyte spores tle habitat during the mid 1990s (see Bishop and and was dated (AMS) at 47,620 ± 2500 yr b.p. Meyer, chap. 14: table 14.2), a technique adapted Bartholomew et al. (2007) have studied joint for application to all Georgia barrier islands in a orientations on the Georgia coast and Rich and continuing longitudinal study by the Coastal Re- Bishop (personal commun.) examined possible sources Division of Georgia Department of Natu- joints at Yellow Banks Bluff in September 2006. ral Resources (Dodd and MacKinnon, 2006). In During the 1980s and 1990s, the University of 2006, investigators working on St. Catherines Pittsburgh group also conducted extensive geo- Island were invited to report surf conditions, rip logical research (largely paleoecological and sedi- currents, and erosional beach conditions during mentological) on St. Catherines Island, including storms (Davis and Dolan, 1993) to NOAA Coast- the work done by K. Beratan, R.M. Busch, J.D. al Services Center, in Charleston, S.C. Donahue, N.J. DeLillo, J.F. Fierstien, S.K. Ken- nedy, D.M. Linsley, P. Mannion-Rowe, R. Pinko- Sea Level History ski, J.E. Pottinger, J.C. Rollins, H.B. Rollins, B.L. of St. Catherines Island Sherrod, C. Venn, F. Vento, and R.R. West. St. Catherines Island is comprised of older Silver Bluff Pleistocene sediment forming the RECENT SEA LEVEL RISE western high-standing core of the island (the is- at St. Catherines Island land core) surrounded by younger, low-standing Holocene accretionary terrains (see fig. 3.2). Sea level rise is occurring at the same time These major sedimentary packages are separated that we have dammed streams, dredged the Sa- by a series of scarps, bluffs, or other dichotomous vannah Ship Channel, and interrupted the south- boundaries (fig. 3.2). These ancient and modern ward flow of sand along the coast of Georgia. erosional boundaries and sedimentary packages Consequently, there is a deficiency in sand sup- provide evidence of episodic erosion and de- ply, placing many islands under increasingly position (see chap. 14: fig. 14.11) as the island erosional conditions, especially St. Catherines changed through time in response to fluctuating Island, which has no significant local influx of sea level, changes in sedimentation rates, and 84 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 crustal movements (Johnson et al., 1974). Such Bishop et al. (2007) presented a comprehen- changes are normal geological phenomena, but sive analysis of the geology of St. Catherines Is- when set in motion with anthropogenic causes or land within the context of a Southeastern section effects are unique and sometimes dramatic. meeting of the Geological Society of American

N N

W E W E

Holocene Accretional S Holocene Accretional S Terrains Terrains

Northwest Engineers Northwest Engineers Scarp St. Scarp Scarp St. Scarp Catherines Catherines Scarp Scarp

North North Beach Beach

Walburg Scarp Walburg Scarp

Seaside Seaside Inlet Inlet

Pleistocenecentral (older) depression Middle Pleistocenecentral (older) depression Middle

Pleistocene (younger) Beach Pleistocene (younger) Beach

King New Ground Scarp McQueen King New Ground Scarp Wamassee Wamassee McQueen Inlet Inlet Scarp Scarp

Scarp Scarp

Atlantic Atlantic Back Creek Ocean Back Creek Ocean

South South Beach Beach

Holocene Accretional Terrains Holocene Accretional Terrains

0 Miles 1 0 Kilometers 2

Fig. 3.2. Geomorphology sketch map (left) and 2006 color imagery (right) of St. Catherines Island (after Bishop et al., 2007, and Linsley, Bishop, and Rollins, 2008) showing major scarps and depressions. 2011 Geoarchaeological Research at St. Catherines Island 85

(SEGSA) Technical Session and Fieldtrip. The generally become progressively younger seaward basic geomorphology of St. Catherines was pre- and southeasterly and can be seen on aerial pho- sented within hypotheses of the formation of the tographs and orthophotomaps to possess discrete, island’s foundation and geological history. This often dichotomous boundaries (figs. 3.2 and 3.5 work brought together aspects of geology, ecol- and Bishop and Meyer, chap. 14: fig. 14.6). These ogy, archaeology, and history. packages of sediment represent rapid periods of Reitz et al. (2008), Linsley, Bishop, and Rol- accretionary activity during the Holocene punctu- lins (2008), and Thomas (2008) examined as- ated by periods of erosional activity giving rise to pects of the natural history and evolution of St. “sedimentary accretionary terrains,” which may Catherines Island, which, along with Bishop’s be chronologically sequenced based on their po- fieldnotes, forms much of the observational data sition by using crosscutting relationships at their reported herein. The publication of Native Ameri- boundaries, and possibly by archaeological dating can Landscapes of St. Catherines Island (Thom- (Thomas, 2008). The pattern of these accretion- as, 2008) provided the impetus for development ary terrains records the depositional history of the of the Caldwell IV Conference stressing the geo- Holocene part of the barrier island in their pres- archaeology of St. Catherines Island, which we ence, sequence, and distribution. Thus the Holo- hope will expand our knowledge by inclusion of cene sedimentary accretional terrains provide a new ideas from the usual suspects and new input powerful tool to decipher the Holocene history of and testing by new research colleagues. the entire island, as well as the record of sea level St. Catherines Island (see fig. 3.1), consists of fluctuation, rate of sediment supply, and shoreline two distinctive entities defined by variations in fluctuation in the recent past. This, in turn, allows topographic relief, or “texture” (fig. 3.2), a high- sedimentological prediction of the near future as standing, relatively featureless central area to the sea level rises due to global warming. west and northwest and a low-standing, highly textured fringing area to the east and southeast Boundaries (Bishop, 1990; Bishop et al. 2007: fig. 3.40; Reitz In general, the boundary between the island et al., 2008). The high-standing area, the island core facies and the Holocene sedimentary accre- core, is characterized by an elevation of approxi- tionary terrains is demonstrably erosional (see mately 5 m, a mature mixed, deciduous-pine for- fig. 3.2). At both ends of the island the low-stand- est, former agricultural fields in various degrees ing linear Holocene beach ridge systems stand of succession, and a robust archaeological re- in marked contrast to the higher, less-textured cord. The high-standing portion of the island is Pleistocene core (figs. 3.3 and 3.4). These ero- mapped as Pleistocene Silver Bluff facies on the sional boundaries represent surfaces formed by geologic map of Georgia and the low-standing episodes of erosion due either to eustatic sea lev- area is mapped as Holocene (Pickering and Mur- el changes, pulses of erosion and sedimentation, ray, 1976). The low-standing accretionary terrain or lateral migration of subtidal erosional environ- is characterized by an elevation of ~1 m, alternat- ments such as channels of sounds and meandering ing beach-dune ridge systems and swales, and in- tidal rivers. Each erosional event is identifiable in tervening freshwater ponds or tidal creek-marsh map view by dichotomous beach ridge orienta- meadows. The ridges are often forested by vari- tions and often by small differences in elevation ous trees, which are rather distinctly distributed of adjacent sequences of sediment, and differenc- along specific ridge systems (Coile and Jones, es in vegetation or archaeological age (fig. 3.3). 1988) and are dominated by cabbage palm, hick- Each erosional event may also be identifiable in ory, pine, or live oak. stratigraphic sections as a minor unconformable The central, high-standing island core is sequence or diastem that dips either seaward or flanked by low-standing Holocene sediments at into adjacent areas that had lowered base levels the north and south ends of the island and along (fig. 3.4). When such erosional surfaces are de- the oceanic eastern margin. The eastern margin stroyed by subsequent erosional events, their for- is characterized by broad marsh meadows de- mer existence and location become conjectural. veloped behind long sand spits at Seaside Inlet, Each identifiable boundary has been given a local Middle Beach, and McQueen Inlet. The southeast name to expedite communication in this chapter end of St. Catherines Island consists of a sequence (Bishop et al., 2007). of approximately 22 beach ridge systems, which The boundary between the island core facies 86 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 and the Holocene sedimentary accretionary ter- several times during the Holocene, the sound rains on the east side of the island consists of channels have migrated dramatically and some- three major arcuate erosional scarp systems (fig. times rapidly, significantly eroding older - por 3.2). The oldest scarp, the King New Ground tions of the Holocene part of the island that sub- Scarp, runs from the south end of Yellow Banks sequently accreted as the sound channel migrated Bluff, the Silver Bluff (Hails and Hoyt, 1968) back in the opposite direction. This hypothesis erosional bluff on North Beach, to Cracker Tom would suggest that the sounds and Pleistocene Creek behind South Beach. Back Creek Scarp cores of barrier islands on the Georgia coast are can be traced from Cracker Tom Causeway to relatively stable features, but that the uncon- the south end of the island core. St. Catherines strained mouths of sounds act somewhat like Scarp is herein restricted to the northernmost part loose fire hoses, whipping rapidly back and forth of the island, essentially fronting Yellow Banks along the coastline, giving rise to the erosional Bluff. King New Ground Scarp is arcuate and boundaries and sedimentary packages of the ac- brings the Silver Bluff core facies into contact cretional terrains. If this is the case, the relative with marsh meadows, which lie topographical- ages of surface exposures and stratigraphic facies ly 3–5 m (10–15 ft) lower than the Silver Bluff tracts ought to confirm the hypothesis. Chowns core. This emarginate boundary, developed on et al. (2008) have postulated the rapid relocation the King New Ground Scarp as former meanders of fluvial-dominated sounds as a mechanism for of the tidal creeks lying within the seaside marsh forming the northeastern islands in island dou- meadows, cuts into the adjacent island core (fig. blets (see chap. 9). 3.4). From Cracker Tom Causeway to the south is Core samples obtained by vibracoring during a second eastward-dipping erosional scarp, Back the fall of 1989 and the spring of 1990 demon- Creek Scarp, whose presence and age relative to strated the stark contrast between the island core the King New Ground Scarp are indicated by the facies and the Holocene sedimentary accretion- presence of a straight scarp forming the east edge ary terrains. The island core facies exposed at of the ancient island flanked to the east by depo- Picnic Bluff and in a core recovered at Mission sitional ridges and swales. Santa Catalina de Guale both consist of rela- The sedimentary accretionary terrains at both tively homogeneous sands that has been deeply ends of the island that are oriented perpendicu- leached and presumably homogenized by in- lar to the northeast-southwest trending beaches tensive bioturbation by root growth and other represent former channel margin sediments of organic processes. The core recovered from the St. Catherines Sound and Sapelo Sound. Because Mission Santa Catalina de Guale site penetrated these areas were formed by lateral migration of clean, light tan, quartz sand at approximately 60 sound systems relatively independent of eustatic cm and terminated at 3.23 m in an organic-rich, change, the lateral erosional and depositional chocolate brown, humic sand. The only evidence events at opposite ends of the island are thought of sedimentary structures was a few sparse root to have been relatively independent of one an- casts and mottling at 2.10 m. Four cores were other. Eustatic events, however, should be cor- taken from the Holocene sedimentary accre- relatable in a large sense as they are drivers of tionary terrains on Cracker Tom Causeway, an gradients and channel morphology. intermediate-age, east-west oriented terrain (see At the north end of the island we now recognize chap. 10: fig. 10.4), and from the northern end two additional scarps, Northwest Scarp and Engi- of Beach Pond, immediately behind the pres- neers Scarp (fig. 3.2). Northwest Scarp is a short ent beach ridge system. All four cores show re- scarp that strikes northeast from Walburg Scarp markable vertical sequences of sediment derived forming the boundary between the island core and from deposition in coastal environments similar Northwest Marsh. Engineers Scarp is the erosional to those nearby. Three of the four terminate in boundary between the island core and the northern deeply colored sediments at approximately 5.0 m Holocene accretionary terrains. Engineers Scarp below current high marsh level. One of these, the is truncated to the east by St. Catherines Scarp, a Cracker Tom Bridge core (GAB 9005051), termi- former sound margin of St. Catherines Sound, and nates in the upper 14 cm of an unknown thickness truncated to the west by Walburg Scarp. of deep blackish-brown, dense peat that has been The number, extent, and variable orientation dated by 14C methods (figs. 10.2D and 10.7). The of Holocene accretionary terrains indicates that surface, represented by deeply colored terrestrial 2011 Geoarchaeological Research at St. Catherines Island 87 sediments at approximately 5.0 m in cores taken Guale Island hypothesis (Bishop, 1990) allows a along Cracker Tom Causeway (see chap. 10: figs. framework for evolution of St. Catherines Island 10.6 and 10.8), is considered to be the contact to be set and tested, although absolute timing of between the Holocene and the top of the Pleis- some events forming the island core remains un- tocene as well as the disconformity representing known (fig. 3.3). Back Creek Scarp in the subsurface. Relict marsh exposed near the southern tip of Constraints on Historical Speculation St. Catherines Island constrains the age of south- The reconstruction of the geological history ward accretion of the Holocene terrains by repre- of St. Catherines Island remains partly specula- senting the youngest preserved, datable lithologic tive at this time because subsurface data are only sequence of marsh deposited before the current now being systematically gathered and analyzed erosional cycle began. This marsh mud contains in the context of surface geology, geomorphol- a detrital peat layer near its surface, overlain by ogy, and archaeology. The major constraints on rooted marsh muds with in situ articulated and dis- such speculation include the lithologies, orienta- articulated shells of the marsh-dwelling bivalves tion, and surficial textures of different parts of the Mercenaria, Geukensia, and Crassostrea as well island; presence of major erosional boundaries as as the marsh-dwelling gastropod Littoraria. The indicated by differing orientations of surficial tex- bedded peat, a piece of wood included in the mud tures and relative ages of sedimentary accretional beneath the peat, and shells of Crassostrea, Geu- terrains; relative dating derived from crosscutting kensia, and Mercenaria from above the peat were relationships; archaeological dating; and absolute sampled and are being dated by 14C methods. dates provided by 14C dating. These data, when The presence of extensive seaside marsh integrated with analogous findings from adjacent meadows along the Atlantic margin of St. Cathe- island systems, provide a technique to decipher rines Island, the presence of a significant erosional the major historical events that have shaped St. bluff on North Beach, the presence of thick, rel- Catherines Island. Among these events, the pres- ict marsh muds exposed along stretches of North ence, position, and relative age of major erosional Beach, Middle Beach, and South Beach, and the boundaries seem most liable to provide sound se- abundance of significant heavy mineral placers quencing of past events. Such boundaries clearly all firmly substantiate that not only has there been separate older sedimentary accretional terrains significant erosion along the Atlantic margin of from younger ones. Such erosional boundaries the island, but that there was a formerly signifi- are not only surficial, but also form disconform- cant barrier island to the east of St. Catherines able relationships in stratigraphic sections. Island. Seaside and McQueen marsh meadows In constructing our more recent geological developed as interisland marshes before they en- history of St. Catherines Island, we summarize tered their current erosional phase. The geomor- evidence as we knew it in 2007 and 2008. From phology of the proposed St. Catherines Island that base, we have constructed an interpretation couplet would be analogous to other Georgia Sea of the geological history of the island constrained Island couplets, such as St. Simons–Sea Island or by parameters cited above (fig. 3.3). The interpre- Sapelo–Blackbeard Island (fig. 3.6). This com- tation of the island’s geological history continues parison leads to the conclusion that a seaward to evolve as new data become available. barrier island, named Guale Island (Bishop, 1990; The evidence, as known in 2008, consisted of Bishop et al., 2007; Linsley, Bishop, and Rollins, the following observations: 2008; and Reitz et al., 2008), once existed off the 1. The oldest part of the island, the Pleisto- northeast edge of modern St. Catherines Island. cene Silver Bluff core, consists of homogeneous Guale Island suffered total destruction from ero- sands, nearly lacking surficial textures, and sion by southerly longshore currents entering the deeply leached with concentrations of organic- Georgia Bight from the northeast and is perhaps rich humate at depth with a central depression linked to stream piracy of the and a lineament that may be a boundary be- by the (Chowns et al., 2008) and tween older and younger depositional units in a subsequent decrease in sediment delivery to St. the island core. Catherines Sound and to St. Catherines Island. 2. Major erosional boundaries are identifiable Although still largely speculative, the re- at the edges of the core, including Back Creek cent history of St. Catherines Island built on the Scarp, St. Catherines Scarp, Walburg Scarp, Wa- 88 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94

A B C

Island Core– Older eline Island Core– Pleistocene St. Catherines Older Pleistocene Shoal

nne Shor

ess A

inc r P transport to SW erosion of shorefront

modern modern modern St. Catherines St. Catherines Sapelo Shoal Sapelo Shoal Island St. Catherines Island Island

0 5 KM 0 5 KM 0 3 KM

Old D Island E Island F Core–Older Core–Older Pleistocene Pleistocene Island Core– Island Core– Younger Younger St. CatherinesCreek Pleistocene Pleistocene King New Guale Ground Scarp Island

Zapala St. Catherines Ridge Sound Erosion

Pleistocene Back Creek welding of younger Scarp to shoreline approx. 32 km

modern modern Old Sapelo River St. Catherines St. Catherines Island Island

0 3 KM 0 3 KM 0 3 KM

Fig. 3.3. A geomorphological depiction illustrating one possible scenario for the development and evolution of St. Catherines Island (Bishop et al., 2007): A, St. Catherines shoal at time of deposition of Princess Anne paleoshore; B, formation of initial Silver Bluff Pleistocene core of island; C, erosion of older Pleistocene core results in long, narrow island and adds sediment to the south; D, welding of younger Pleistocene core onto entire length of island; E, meander of Zapala Sound erodes older Pleistocene coupled with development of complex St. Catherines/Guale barrier island doublet; F, Wisconsin low-stand shoreline recedes 32 km east near Gray’s Reef, island is part of low-relief mainland. 2011 Geoarchaeological Research at St. Catherines Island 89

G H I Engineers Scarp Island Core– Island Core– Island Core– Older Older Guale Guale Older Guale Pleistocene Pleistocene Pleistocene Isl. Isl. Isl. Island Core– Island Core– marine Island Core– Younger Younger terrace Younger terrestrial Pleistocene Pleistocene Pleistocene deposit

o SW t t

osion of Guale anspor er tr

older accretional terrains Zapala modern modern Sound modern St. Catherines St. Catherines St. Catherines Island Island Island

0 3 KM 0 3 KM 0 3 KM

northern northeast– J accretional K L continued terrains accretion Island Core– Island Core– Older Older Pleistocene Pleistocene Island Core– Island Core– Younger Younger Pleistocene LEGEND Pleistocene Northwest Scarp St. Caths. Scarp east– Walburg continued Scarp retreat Wamassee Scarp Engineers Scarp south– continued King New erosion Ground Scarp Back Creek 0 3 KM 0 3 KM Scarp 0 3 KM

Fig. 3.3 (continued). G, Sea level rises, Guale Island erodes and sand is transferred southward forming the older accretional terrains; H, destruction of Guale with transfer of sand forms barriers protecting Seaside and McQueen marshes as North Beach is exposed to the ocean and erosion forms a marine terrace as Zapala Sound migrates north, truncating older accretional terrains and forming terrain #6 of fig. 3.5B; I, sand transfer south from Guale continues building southern accretional terrains, meandering tidal creeks erode emarginations into King New Ground Scarp, and blowing sand buries the terrace; J, present configuration of the island as Native Americans found it; K, present-day island with major scarps overlain, and L, future configuration of the island using current accretional/erosional areas. 90 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 massee Scarp, and the modern shoreline, which is Bishop, 1993; Darrell, Brannen, and Bishop, undergoing erosion over most of its length. 1993; Bishop and Marsh, 1999a; Bishop et 3. The King New Ground Scarp, forming the al., in press). In the normal operations of the north-south boundary between the Silver Bluff St. Catherines Island Sea Turtle Program, op- Core and the Holocene Seaside Marsh Meadow portunities have arisen to make numerous ob- from Picnic Bluff at least to Cracker Tom Cause- servations and deductions, and to participate in way has been partially obscured by ancient me- application of new technologies to investigate ander scars of tidal creeks where they impinged the formation of the island. Some observations on the island core. were directly associated with the St. Catherines 4. The newer Holocene part of the island Island Sea Turtle Program and others were made consists of highly variable, textured sequences as asides to this program. A review of the timing of sediment of varying type, often exhibiting di- of these observations (Bishop: personal records, chotomous boundaries and orientations of beach field notebooks) was made to attempt to define ridge systems, which can be ordered in relative the flow of the observations, deductions, and ap- depositional sequence. plications … and what each had to do with our 5. Relict marsh is currently exposed along evolving concept of St. Catherines Island. all beaches, especially the south end of North During the initial phases of research on St. Beach, nearly all of Middle Beach, and the north Catherines by Bishop (1986–1990) and during and south ends of South Beach (Morris and Rol- initialization of the St. Catherines Island Sea Tur- lins, 1977). tle Program (1990–present), it became obvious 6. The Seaside Marsh muds exposed behind that unknown circumstances were negatively af- Black Hammock Spit on the south end of North fecting hatching success of in situ sea turtle eggs Beach are at least 9.14 m (30 ft) thick in places. in nests along part of North Beach between Sand 7. Evidence of ancient tidal creeks is pre- Pit Road entrance and Yellow Banks Bluff. It was served as two oxbows at the north end of Picnic postulated that ground water conditions were re- Bluff and as sand-filled meanders near the south sponsible for the poor hatching conditions, so four end of North Beach. standpipes were jet-drilled into the accretional 8. The concentration of heavy mineral placers terrace from the backbeach toward the southwest on the beaches is anomalously high (Jack Reyn- into an ancient oxbow. Two oxbows were pres- olds, personal commun.), indicating winnowing ent at that time, one between Picnic Bluff and the of a considerable amount of sediment from erod- major part of Yellow Banks Bluff and one to the ing shorelines. north, just south of Sand Pit Road (into which 9. Vibracores through the Holocene accretion- the southwestern-most standpipe was inserted). ary terrains penetrate multiple progradational se- During the drilling process, and as indicated in a quences separated by erosional surfaces. vibracore taken by Bran Potter in his island ecol- 10. Vibracores in the sedimentary accretional ogy class (Bishop et al., chap 10: fig. 10.3), a peat terrains on Cracker Tom Causeway and on the and underlying marsh mud were encountered at a west side of the island at St. Catherines Shell depth of about 2 m on the front side of the island. Ring penetrate deeply colored peats and marsh Another core, further south near the south end of deposits near the Holocene-Pleistocene contact at Yellow Banks Bluff, replicated this anomalous approximately 5 m. condition in an area in which we expected to en- 11. 14C dates, where possible, allow surfaces counter mostly sandy sediment. and lithosomes to be dated. On August 26, 1996, colleague Nancy Marsh 12. Archaeological data could indicate when reported the presence of exposed ghost shrimp a surface was available to human utilization, a burrows eroding along the front of North Beach. minimum age of formation. Upon checking, the knobby, mud-lined burrows proved indeed to be the lower part of the bur- Sea Turtles, Sea LevelS, rows, a horizontal burrow maze that forms at 2–5 and Geology m below the midbeach level. This elevation of the lower part of burrows indicated that, at the Observations that are anomalous or repre- time the burrows were active, sea level was high- sent a misfit to our worldview form the basis er than it is now by about 2–5 m or so. This de- of the search for new knowledge (Brannen and duction, along with continuing erosion of Yellow 2011 Geoarchaeological Research at St. Catherines Island 91

Banks Bluff, led to a close examination of the ples, Ron West, Dave Linsley, and Blaine Cecil), bluff (often cited as the only exposed and erod- and then by Bran Potter and Tim Keith-Lucas and ing Pleistocene on the Georgia coast). On July the island ecology program (~1990). This led to 17, 2001, Bishop measured a detailed section at the testing of stratigraphic hypotheses and the de- Yellow Banks Bluff and depicted small, unlined velopment of several transects to define bound- burrows throughout half of the section and a zone aries constraining the island. On May 25, 1990, of truncated burrows in a horizontally laminated Bishop met with Bud Rollins and Dave Thomas humate layer about 20 cm below an intensively and presented a hypothesized geological history burrowed surface 3.6 m below the top of the of St. Catherines Island, the trigger for much sub- bluff (fig. 3.4). In this section, a possible ghost sequent work on its evolution. shrimp burrow was identified about 1.7 m below Data from isolated vibracores provided in- the modern surface and charcoal and a whelk triguing evidence of the marine origin of the shell were collected at 1.39 m below the top of island’s foundation. This evidence included the the bluff. Bishop’s field notes record attempts presence of lined burrows known as Ophiomor- to resolve the anomalously high occurrence of a pha nodosa, the burrows of the Carolinian ghost ghost shrimp burrow, at least two possible dis- shrimp, Callichirus major (Say, 1817–1818) conformities, and the presence of charcoal in the (Bishop and Bishop, 1992; Bishop and Williams, “Pleistocene” sediment of Yellow Banks Bluff. 2005), interlaminated layers of heavy minerals Later that summer, on September 29, 2001, a and quartz sand indicative of the backbeach fa- “short section” of the bluff was measured and cies, and occasional shells of marine molluscs, joints were recognized, being marked by nearly comparisons with uplifted barrier islands (How- “vertical burrow clumps cemented by humate ard and Scott, 1983; Pirkle et al., 2007) and pa- along joints.” One joint set was oriented paral- lynological data (Booth and Rich, 1999). lel to the bearing of North Beach and inclined In preparation for publication of Native steeply toward the ocean, forming the surface American Landscapes of St. Catherines Island along which the bluff scarp formed. Charcoal (Thomas, 2008), several collaborative chapters was seen and sampled 1.37 cm from the top of on island natural history and geology were writ- the bluff (but not dated). ten (chap 3: Stratigraphy and Geologic Evolution During the 1990s, sea turtle skeletons were of St. Catherines Island [Linsley, Bishop, and buried in the sand dunes behind North Beach to Rollins, 2008] and chap 5: A Brief Natural His- recover osteological specimens from dead and tory of St. Catherines Island [Reitz et al., 2008]). stranded vertebrates, especially sea turtles and Linsley, Bishop, and Rollins (2008), describe birds. Burial in the backbeach sand allows the several vibracore transects that were construct- soft parts to decompose and the skeleton to be ed along the margins of St. Catherines Island, collected as a virtually clean specimen in about a including the Cracker Tom Transect drilled by year. Several skeletons were buried in the highest Bishop, Rich, and Hayes and described as single “clean” sand at the point where Sand Pit Road sections in Booth, Rich, and Bishop (1999) and drops off the “Pleistocene” older part of the is- as transects in Bishop et al. (2007: fig. 36), and land. Excavation of a shallow burial trench about Linsley, Bishop, and Rollins (2008: fig. 3.3.9). In 1 m deep (in which a sea turtle would be buried) addition, two new transects across Yellow Banks exposed interlaminated horizontal heavy mineral Bluff (transect A–A′) and the north end of Sea- and quartz horizontal sand layers at this position. side Spit (transect B–B′) were described in Lin- The horizontally laminated sand was recognized sley, Bishop, and Rollins (2008). Subsequently as a backbeach facies at an anomalously high el- the St. Catherines Shell Ring transect near Long evation (Wharton, 1978; Howard and Frey. 1980; Field was drilled by Keith-Lucas, Potter, Bishop, Howard and Scott, 1983; Frey and Howard, and Thomas’ AMNH archaeology crew (Bishop 1988; Bishop et al., 2007: figs. 11–12; Milliken, et al., 2007: fig. 3.6I; chap 10, this volume). Anderson, and Rodriguez, 2008). These collaborative publications sparked The acquisition of a vibracore rig by Rich and further collaborative research by the AMNH ar- Bishop in 1989 (Bishop et al., 2007) initiated a chaeology program and brought the vibracore program of vibracoring on St. Catherines Island, technique into their archaeological repertoire on which was rapidly enlarged by Bud Rollins and St. Catherines Island in 2005 as the island ecol- his drill team on 8/17/90 (Rich Busch, Chris Ma- ogy program class drilled three vibracores in the 92 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94

N Overview Map c North Beach Oxbow Transect W E west east S a c b Yellow North accretional ridges d Banks Oxbow (Holocene) WT MHW WalburgWalburg Creek Creek beach sand peat MLW King New Ground e marsh mud Emarginate Scarp channel lag Seaside InletInlet d Yellow Banks Blu Transect west east 35 m of blu retreat 1971–1996 (25 years) 6 McQueen InletInlet 5 St.St. CatherinesCatherines IslandIsland 4 3.69 m range of sea KingKing New New Ground Ground (KNG) (KNG) Scarp Scarp 3 level from horizontal burrow maze Yellow MHW 2 Banks 1 beach sand 0 MLW meters marsh mud above MLW

e Seaside Marsh/Spit Transect west east Fish marsh mud Former King New exposed 6 Creek Seaside Tributary Ground Scarp (?) Seaside Spit Spit 4 Fish Fish washover (2006) (1974) Creek Silver Creek fans sand 2 Blu Tributary (Pleist.) marsh mud MSL tidal channel and -2 King New Ground channel lag facies (Holocene) Emarginate Scarp King New Ground Scarp -4 Horiz. Scale (m) 0 100 Pleistocene

Fig. 3.4. Evidence of a mid-Holocene sea level highstand on the north end of St. Catherines Island. Subsurface depth and sediment type adapted from Linsley, Bishop, and Rollins (2008), position of scarps from surface geology/aerial imagery, subsurface relationship of scarps suggested from surface geology, limited core data, and modern depositional environments. 2011 Geoarchaeological Research at St. Catherines Island 93

A St. Catherines N W E 0 500 1000 1500 FT Sound S 0 150 300 450 M

Holocene northern Holocene northeastern

accretional Scarp Catherines St. northeastern

Walburg Creek terrainsterrains accretional Holocene terrainsterrains Walburg Creek Engineers Scarp accretional terrainsterrains

Atlantic Pleistocene Ocean islandisland corecore (IC)(IC) Northwest Scarp

Walburg Scarp

Pleistocene B 2 N islandisland South Newport W Scarp 9 E core (IC) 1 Johnson Creek 3 S River Back Creek 4 5 7 8 10 6 11 Holocene southern 12 13 accretional Atlantic terrainsterrains Ocean 14

15

16

17 0 500 1000 1500 M 18 19 20 Imagery Source: USDA NAIP 1999 Fig. 3.5. Depositional components of St. Catherines Island include a Pleistocene Island Core (IC), erosional scarps, accretional terrains composed of progressively younger beach ridge systems, and modern dune fields: A, north end of St. Catherines Island showing the Pleistocene Island Core (IC), scarps, and Holocene accretional terrains deposited as St. Catherines Sound migrated back and forth; B, south end of St. Catherines showing progressively younger Holocene accretional terrains (1 is the oldest) toward east and southeast. 94 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 marsh north of the newly documented St. Cath- Observational Results erines Shell Ring in Meeting House Field. Bish- op and the AMNH archaeologists, led by Matt The sum of this collaborative research led Sanger and Elliot Blair, extended this transect us to question the timing and effects of the by drilling four new cores connecting the island Pleistocene sea level rise at St. Catherines Is- ecology program with subsequent vibracoring land and attempt to define hypotheses to answer within the shell ring to further define geophysi- these questions. The geological record has left cal anomalies located by magnetic, resistivity, its imprint as the stratigraphy and geometry of and ground penetrating radar (GPR) geophysics the island (Demarest and Kraft, 1987). Specifi- (Bishop et al., chap. 14). cally, past observations used in 2007 and 2008 In 2005, Kelly Vance began the geological now need to be integrated with new observa- profiling of St. Catherines Island using anew tional evidence on the evolution of St. Cathe- GSU ground penetrating radar unit. Profiles were rines Island, including: (1) a central depression done along existing roads and on the beach (see apparently separating an older terrain from a Vance et al., chap. 11). Several anomalies includ- younger, slightly higher, terrain, (2) horizontal ing the Gator Pond collapse site at the north end backbeach laminations in State Road vibracore of the island, the Central Depression, and a dis- (chap. 10: fig. 10.5) 2.5 to 2.9 m in the center of conformity near South End Settlement were iden- the central core on State Road Pond, (3) hori- tified by this reconnaissance work, and additional zontal backbeach laminations at an anomalously attention was focused on the “Pleistocene” bluff high elevation on Sand Pit Road as it leaves the (which was beginning to be thought of as possi- central core, (4) the highly eroded backbeach bly a Pleistocene or Holocene dune fill of a Pleis- scarp hosting large meander scars, (5) the two tocene erosional terrace). In 2006, Tony Martin oxbows cut into St. Catherines Scarp north of, and Andy Rindsberg visited St. Catherines Island and within, Yellow Banks Bluff, (6) horizontal to assess Neogene traces and especially the bur- burrow mazes of ghost shrimp in front of and rows present in Yellow Banks Bluff (Martin and covered by Yellow Banks Bluff in the recent Rindsberg, 2008; chap. 5, this volume), which past, (7) the erosional terrace or washover fan they ascribed to fiddler crabs at the base and bur- (Martin and Rindsberg, chap. 5) at 1.5 m above rowing by beetles (perhaps cicadas) throughout high tide line at Yellow Banks Bluff burrowed the sediment above. The age of these dune sedi- by fiddler crabs, (8) covering of the erosional ments has now been dated by absolute methods terrace by apparent terrestrial dune sediment (see also chap. 4). burrowed by beetles and weathered into several The geological research on St. Catherines paleosols (Vento et al., 2008; and chap. 4, this Island and its implications for sea level change volume) with possible associated cultural fea- in the western North Atlantic along the coast of tures, (9) GPR data of Kelly Vance document- Georgia resulted in the organization of a techni- ing a disconformity under South End Settlement cal session at the Savannah meeting of the South- and a collapse structure under the north end Ga- eastern Section of Geological Society of America tor Pond, (10) Pleistocene fern-peat beds flank- (Bishop, Vance, and Meyer, 2007) and a SEGSA ing the eastern and western sides of the island, fieldtrip to St. Catherines Island (Bishop et al., (11) the sequencing of the accretional terrains 2007). in the Holocene, (12) the development of two In 2008, Tim Chowns, who had just attended northwest trending dunes on top of the east-west the SEGSA St. Catherines fieldtrip, published a terrains on the north end (Brian Meyer, personal paper hypothesizing the rapid movement of the commun., 2009), and (13) an erosional event Georgia sounds as they responded to rising and subsequent to that of 1867 (chap. 8) on the north falling sea level (Chowns et al., 2008). He sug- end forming a new accretional terrain now host- gested a vibracoring program on St. Catherines ing the Sand Pit Road sea turtle rookery (chap. Island and Blackbeard Island (chap. 9, this vol- 14: fig. 14.9A and table 14.2). ume) to define what he envisioned as cutoff spits The observations listed above, and those made forming Blackbeard Island on the Sapelo/Black- by others, must be explained by scenarios of the beard Island doublet and the hypothesized Guale development of St. Catherines Island, such as Island on the St. Catherines/Guale Island doublet those by Booth et al. (1999), Bishop et al. (2007), (Bishop et al., 2007: fig. 6). and Rollins, Prezant, and Toll (2008). As visual 2011 Geoarchaeological Research at St. Catherines Island 95 learners, and thinkers, Bishop and Meyer have Silver Bluff shoreline. Sea level then dropped, attempted (several times) to construct a sequence and this shoal developed into St. Catherines Is- of evolutionary stages for the development of St. land, which probably developed a fairly level Catherines Island (as in fig. 3.3). surface by eolian transport of beach sediment. Any island geological history must also be Sea level then rose slightly against this old island directly tied to the present geomorphology be- and eroded its front and back sides, forming a cause the topography, bathymetry, scarps, palyn- narrow strip-like older Pleistocene barrier island, ology, and accretional terrains are the physical then dropped slightly (?), and rose again approxi- record of past island morphology and its contin- mately to its previous level, welding a younger gency (Gould, 1989). However, the position of Pleistocene island onto the older Pleistocene one. these features only constrains the portion of the Rising sea level established a new offshore island island that is still preserved and visible; it can- on the northeast, Guale Island, forming the island not address the portions that formerly existed be- doublet of St. Catherines–Guale Island similar in fore the preserved features formed. This type of size and shape to the seaward islands of extant constrained reconstruction was done by Bishop island doublets, such as Sapelo-Blackbeard (fig. et al., 2007, fig. 3.70, in which the island core 3.6). These islands effectively baffled the front of was cut and pasted into an evolutionary diagram. modern St. Catherines, allowing the formation Thomas, Rollins, and DePratter (2008), in their and continued growth of an inter-island marsh analysis of the shape of St. Catherines, presented that kept pace with rising sea level, eroding new the changing shape of the island on a background scarps along the front of the island (King New of the present island outline, a nice metaphoric Ground and Back Creek erosional scarps). device followed here (and also on previous hand- During the last glaciation, the Wisconsin, sea drawn analyses by us), but strongly enhanced, we level dropped to at least 80 m (262 ft) below its think, by placing the developing island onto the current level, causing the seashore to withdraw far geomorphological map of St. Catherines Island to the east (approximately 32 km east near Gray’s (fig. 3.3). The hypothesized formation of parts Reef) and exposed the upper shelf to colonization of the St. Catherines and Blackbeard islands by by plants and animals and to erosion. St. Cathe- piracy of St. Catherines Sound and Sapelo Sound rines Island (10,000 yr b.p.) would have been part (Chowns and Stogner, 2008; Chowns et al., 2008) of the mainland, a hill-like ridge probably lying is developed elsewhere in this volume (chap. 9). between two valleys occupied by small streams, with Ancient Sapelo and Ancient Medway rivers Constructing a Scenario (Riggs and Cleary, 1993) occupying the courses of Island History of modern Sapelo and St. Catherines sounds. Be- cause of its relief, St. Catherines Ridge would In setting any scenario, we must first deter- have undergone erosion as the streams extended mine what we know and integrate that with what themselves headward, forming small tributary we think we know, then formulate multiple work- gullies cutting into the hill. Plants and animals ing hypotheses to explain them (Chamberlin, would have been free to colonize St. Catherines 1890). We do know that sea level is not static, Island at this time. Fern peat on either side of the and, in fact, has fluctuated significantly through- island (47620 ± 2500 yr b.p., USGS WW1197 and out the last 1.8 million years with the waxing and >44,800 yr b.p. Beta 217823) records the pres- waning of the ice sheets of the Pleistocene, caus- ence of conifer forests with a fern understory on ing sea level to fluctuate approximately 200 m St. Catherines Ridge, as the shoreline lay 50–80 (~600 ft) and currently coming to stand near its km east. As the Wisconsin Ice Sheet reached its median height. maximum extent, sea level was near its minimum, St. Catherines formed sometime prior to about 120 m below current sea level, then began 44,000 yr b.p. as a barrier island of the Silver its rapid rise about 10,500 yr b.p. By about 6020 Bluff paleoshoreline from an older island or shoal yr b.p., the peat at Cracker Tom Marsh (Booth (marked by horizontal backbeach laminations at and Rich, 1999) was covered by saltwater with –2.5 m under State Road Pond (chap. 10: fig. marine shells cast upon the disconformity, or in 10.5). Modern St. Catherines Island originated as the sediment deposited immediately above them. a Silver Bluff barrier island as determined by its Sea level must have continued to rise and started elevation and alignment with other barriers in the to erode Guale Island, passing its sediment as a 96 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 packet or bulge southward along the face of the major erosional scarps, all apparently associated Pleistocene core of St. Catherines Island, first al- with migrations of St. Catherines Sound. The first lowing erosion of a marine terrace exposed at the (Northwest Scarp) eroded the northwest corner base of Yellow Banks Bluff reaching its highest of the island followed by progradation of a nar- level of recent age (1.5 m above current high tide row marsh and a few north-south trending beach level), depositing laminated backbeach sediment ridges. The second erosional event (Engineers at Sand Pit Road and eroding the Yellow Banks Scarp) was a significant southward migration of Bluff marine terrace marked by fiddler crab bur- St. Catherines Sound, which eroded the northern rows near the base of Yellow Banks Bluff. Sea tip of St. Catherines Island back to a significant level then dropped, exposing the St. Catherines degree and then was followed by northward pro- Ebb Delta to wind erosion, building a dune field gradation forming the bulk of the northern tip of along the leading edge of the island and filling modern St. Catherines Island. A third event, dat- the void left by the Yellow Banks Bluff marine ed by historical data at 1867 (Picnic Bluff Scarp) terrace with dunes that were home to cicadas or eroded into the northeastern corner of the island other beetles. Human occupation occurred ap- and is currently rehealing itself by progradation proximately 5000 yr b.p. (Thomas, 2008). to the east, forming the accretional terrain on the The north end of St. Catherines exhibits three northeast shoulder of the island and the Sand Pit

N N

E W E W S Blackbeard S Guale IslandIsland IslandIsland

Sapelo St. Catherines IslandIsland IslandIsland

0 1 2 0 1 2

KM KM

Fig. 3.6. The Sapelo/Blackbeard Island doublet as a modern analog to the hypothesized St. Catherines/ Guale Island doublet. 2011 Geoarchaeological Research at St. Catherines Island 97

Road sea turtle nesting rookery (see chap. 8). The tinue forming today. Drowned beach ridges north St. Catherines Scarp formed when the east face of Cracker Tom Causeway must have formed on of the island was exposed once Guale Island was the rising sea level and are partly drowned by in- destroyed and the northern accretional terrains undation. Another major erosional event, marked had been deposited and now form the arcuate by the Cracker Tom Scarp, occurred cutting an trace of Yellow Banks Bluff. arcuate reentrant into the middle part of the is- Meandering tidal creeks in Seaside Marsh land and forming an oblique edge on Cracker periodically cut meanders into the adjacent, Tom Hammock as well as depositing a prominent high-standing island core forming King New beach ridge extending from the outer edge of a Ground Scarp. A second, or later event, marked feature we call hickory hill hammock northwest- by the Back Creek Scarp occurred, cutting ward into the marsh. Subsequent to this erosional deeply into the southeastern part of the Silver event, the reentrant was rehealed by progradation Bluff core, which was immediately followed of the island to the east, eventually straightening by lateral transport of sediment forming the its front edge along South Beach. The south end oldest preserved sedimentary accretional ter- of the island has progressively prograded south- rain (terrain 1) comprised of beach ridges and ward with only minor erosional fluctuations, in- a long northeasterly trending spit (Gardener’s cluding the one in which we find ourselves today Hammock), which reaches far into the marsh (Griffin and Henry, 1984). north of Cracker Tom Causeway. St. Catherines The presence and chronology of human oc- Sound periodically meandered southward, erod- cupation will indicate the age status of the ac- ed the east-west portion of St. Catherines Scarp, cretionary ridges at the time of immigration (De- and laid down a series of east-west terrains as Pratter and Howard, 1981; Mason, 1993; Thom- it migrated back toward the north. As the sedi- as, 2008). Sediment from the now eroded Guale ment from Guale Island continued to be trans- Island and St. Catherines Ebb Delta continued to ported southward, it accreted to the southeast be transported southward, accreting as a series of of the island as a series of accretional terrains hooklike terrains to the southeast of the island, that were often marked by dune ridge systems forming a series of accretional terrains south and forming parallel to the shorelines. An interval east of Zapala Shoreline. Discrete boundaries and of progradational sedimentation with minor “stranded” marshes attest to fluctuating sedimen- episodes of erosion followed, building the ac- tological conditions as either sea level fluctuated, cretionary terrains (terrains 2–5) exposed on the sediment supply fluctuated, or storms modified inner part of Cracker Tom Causeway. As this the beaches (Langley et al., 2003). Two longitu- transport was occurring, the sound south of St. dinal dune ridges arose northwestward from St. Catherines migrated northward, its north mar- Catherines Bluff covering the east-west terrains gin eroding across the island, forming a sound of the north end of St. Catherines Island (see margin called Zapala Shoreline. A major ero- chap. 8). Attachment of fringing marshes to the sional event, marked by the Zapala Shoreline, west side of St. Catherines Island has since oc- then occurred as Sapelo Sound swept northward curred, although its timing is yet unknown. and completely truncated the south end of St. The position and shape of St. Catherines Is- Catherines along a transverse erosional surface land seems remarkably stable through time, al- running completely across the island from South though its size apparently changes. The surface Settlement to the beach. The channel margin of of St. Catherines Island core is remarkably level Sapelo Sound then began migrating southward, and represents a packet of sediment that overlies leaving behind a series of transverse ridges and and buries the cultural features across the island. swales (preserved as terrains 6–22). The depth of burial of cultural features gives an Meanders formed the two oxbows at Yellow approximate measure of the rate of sedimentation Banks Bluff as sea level was dropping. A perva- across the island. The relative flatness of the is- sive marsh existed on the ocean side of St. Cath- land surface indicates that alluviation is by pro- erines, with meandering tidal creeks (that still cesses that result in relatively even distribution of exist today) and ebb deltas like that at McQueen sediment or redistribution of the sediment by sub- Inlet (Shadroui, 1990). Meander reentrants cut- sequent processes that level it. Along the beaches, ting across Back Creek Scarp into the Pleistocene we have observed windblown sand being eroded core formed as sea level was dropping, and con- from North Beach and/or Yellow Banks Bluff 98 ANTHROPOLOGICAL PAPERS AMERICAN MUSEUM OF NATURAL HISTORY NO . 94 being transported on northeastern winds up the 1. St. Catherines Island remains a dominant- face of Yellow Banks Bluff, accumulating on top ly erosional barrier island, if not the most ero- behind the lip of the bluff. Hayes (1967) studied sional island of the Georgia Golden Isles, truly a the hurricane effects of Hurricane Carla on the sentinel island for the Georgia coast. Texas coast, attributing the most profound ef- 2. Erosion is now occurring on virtually all fects to wind waves and storm surges. Morton, margins of the island. Erosion rates average Gelfenbaum, and Jaffe (2007) summarized the ~2.0 m/yr with most rapid erosion at Middle effects of Hurricane Camille on the Alabama bar- Beach and Seaside Spit and at the south end of rier islands in 1969. In addition to dune and scarp the island (chap. 14: fig. 14.13). erosion commonly attributed as hurricane effects, 3. Coastal erosion can be adequately char- they documented the transfer of that eroded sand acterized and indexed by observational criteria, across the barrier islands as a series of washover including scarping, boneyards, washover and fans, terraces, and sheetwash, often washing and washin fans, exposed relict marsh mud on the leveling the sand 200–300 m behind the eroded beach, and root zones and peat beds exposed scarps and dunes. Scott et al. (2003) present on the beach. evidence for prehistoric hurricanes on the South 4. Ancient scarps, depositional terrains, and Carolina coast. This catastrophic process over dichotomous boundaries can be related to past time becomes a powerful transporting and level- episodes of coastal erosion and accretion. ing process. Of course, once the sand is blown or 5. The evolution of St. Catherines Island washed on top of an island ,it can be redistributed is decipherable, but the accepted scenario will by plant roots, animals, rain sheetwash, and hu- continue to evolve as new observations are man activities … bringing us into the realm of made and more data becomes available. archaeology. 6. The erosional conditions documented at St. Catherines will migrate northward and Conclusions southward on the Georgia coast as sea level continues to rise, sediment continues to be im- The conclusions supported by the data cited pounded, and anthropogenic modification of the herein include: coast continues.